CN111368596A

Movatterモバイル変換

Info

Publication number: CN111368596A
Application number: CN201811599939.9A
Authority: CN
Inventors: 周军; 吴歌; 何笛
Original assignee: Beijing Techshino Technology Co Ltd; Beijing Eyecool Technology Co Ltd
Current assignee: Beijing Techshino Technology Co Ltd; Beijing Eyecool Technology Co Ltd
Priority date: 2018-12-26
Filing date: 2018-12-26
Publication date: 2020-07-03
Anticipated expiration: 2038-12-26
Also published as: CN111368596B

Abstract

Translated fromChinese

本发明公开了一种人脸识别逆光补偿方、装置、计算机可读存储介质及设备，属于人脸识别领域。该方法包括：在当前逆光补偿参数下采集人脸图像，并在采集到的人脸图像上定位出人脸区域；计算人脸区域的多个指标并确定多个指标的权重，所述多个指标包括人脸区域灰度指标和/或人脸区域频率域指标；根据所述人脸区域的多个指标以及多个指标的权重计算调整参数；判断调整参数是否在预设的阈值范围内，若是，不调整当前逆光补偿参数，否则，调整当前逆光补偿参数。本发明适用于室内暗光、室内强逆光、室外强逆光、夜晚室外环境下的自动补光，能够自适应调节逆光补偿参数，解决暗光及逆光环境下人脸图像过曝或过暗从而使得人脸面部信息缺失的问题。

The invention discloses a face recognition backlight compensator, a device, a computer-readable storage medium and equipment, and belongs to the field of face recognition. The method includes: collecting a face image under current backlight compensation parameters, and locating a face area on the collected face image; calculating multiple indicators of the face area and determining the weights of the multiple indicators, the multiple The indicators include the grayscale indicators of the face area and/or the frequency domain indicators of the face area; the adjustment parameters are calculated according to the multiple indicators of the face area and the weights of the multiple indicators; it is judged whether the adjustment parameters are within the preset threshold range, If so, do not adjust the current backlight compensation parameters, otherwise, adjust the current backlight compensation parameters. The invention is suitable for automatic fill light in indoor dark light, indoor strong backlight, outdoor strong backlight, and outdoor environment at night, and can adaptively adjust the backlight compensation parameters, so as to solve the problem of overexposure or too dark of the face image in the dark light and backlight environment, so that the The problem of missing face information.

Description

Translated fromChinese

人脸识别逆光补偿方法、装置、可读存储介质及设备Face recognition backlight compensation method, device, readable storage medium and device

技术领域technical field

本发明涉及人脸识别领域，特别是指一种人脸识别逆光补偿方法、装置、计算机可读存储介质及设备。The present invention relates to the field of face recognition, in particular to a method, device, computer-readable storage medium and equipment for backlight compensation of face recognition.

背景技术Background technique

人脸识别，是基于人的脸部特征信息进行身份识别的一种生物识别技术，用摄像机和摄像头采集含有人脸的图像或视频流，并在图像中检测和跟踪人脸，进而对检测到的人脸进行脸部的一系列相关技术，目前，人脸识别技术已经被广泛应用于安防、金融、公共安全等领域。Face recognition is a biometric recognition technology based on human facial feature information. It uses cameras and cameras to collect images or video streams containing human faces, and detects and tracks human faces in the images. Face recognition technology has been widely used in security, finance, public safety and other fields.

在进行人脸识别时，摄像机或摄像头拍摄图像质量对人脸识别的效率和准确率非常重要，图像质量的好坏与环境光照条件紧密相关。现有技术中一般采用固定逆光参数的近红外摄像头，当光照充足时，可以拍摄出明亮的、清晰图像，当光照不足时，比如夜间，拍摄图像亮度不够，且存在大量图像噪声，无法清晰成像，造成人脸识别时漏检率、拒识率高，甚至无法实现人脸识别。为了解决这种现象，现有技术中的一种方法是采用白光进行逆光补偿，由于人眼可以感知白光，在采用大功率白光进行逆光补偿时，会使人炫目，特别是在交通路口，很容易引发交通事故，若降低补光功率，又无法清晰成像，仍然达到不到人脸识别的要求。When performing face recognition, the quality of images captured by cameras or cameras is very important to the efficiency and accuracy of face recognition, and the quality of the images is closely related to the ambient lighting conditions. In the prior art, a near-infrared camera with fixed backlight parameters is generally used. When the illumination is sufficient, a bright and clear image can be captured. When the illumination is insufficient, such as at night, the brightness of the captured image is insufficient, and there is a large amount of image noise, which cannot be clearly imaged. , resulting in a high rate of missed detection and rejection during face recognition, and even unable to achieve face recognition. In order to solve this phenomenon, a method in the prior art is to use white light for backlight compensation. Since human eyes can perceive white light, when high-power white light is used for backlight compensation, it will make people dazzled, especially at traffic intersections. It is easy to cause traffic accidents. If the fill light power is reduced, the image cannot be clearly imaged, and the requirements of face recognition are still not met.

发明内容SUMMARY OF THE INVENTION

为解决上述技术问题，本发明提供一种人脸识别逆光补偿方法、装置、可读存储介质及设备，本发明适用于室内暗光、室内强逆光、室外强逆光、夜晚室外环境下的自动补光，能够自适应调节逆光补偿参数，解决暗光及逆光环境下人脸图像过曝或过暗从而使得人脸面部信息缺失的问题。In order to solve the above technical problems, the present invention provides a face recognition backlight compensation method, device, readable storage medium and equipment. The present invention is suitable for automatic compensation in indoor dark light, indoor strong backlight, outdoor strong backlight, and outdoor environment at night. It can adaptively adjust the backlight compensation parameters to solve the problem that the face image is overexposed or too dark in the dark light and backlight environment, resulting in the lack of face information.

本发明提供技术方案如下：The present invention provides technical solutions as follows:

第一方面，本发明提供一种人脸识别逆光补偿方法，所述方法包括：In a first aspect, the present invention provides a face recognition backlight compensation method, the method comprising:

在当前逆光补偿参数下采集人脸图像，并在采集到的人脸图像上定位出人脸区域；Collect a face image under the current backlight compensation parameters, and locate the face area on the collected face image;

计算人脸区域的多个指标并确定多个指标的权重，所述多个指标包括人脸区域灰度指标和/或人脸区域频率域指标；Calculate multiple indices of the face area and determine the weights of the multiple indices, the multiple indices include the face area grayscale index and/or the face area frequency domain index;

根据所述人脸区域的多个指标以及多个指标的权重计算调整参数；Calculate the adjustment parameter according to the multiple indicators of the face area and the weights of the multiple indicators;

判断调整参数是否在预设的阈值范围内，若是，不调整当前逆光补偿参数，否则，调整当前逆光补偿参数。It is judged whether the adjustment parameters are within the preset threshold range, if so, do not adjust the current backlight compensation parameters, otherwise, adjust the current backlight compensation parameters.

进一步的，在当前逆光补偿参数下采集到的人脸图像为多帧，所述方法还包括：Further, the face images collected under the current backlight compensation parameters are multiple frames, and the method further includes:

根据采集到的多帧人脸图像的变化判断人脸图像是否稳定，若是，执行所述计算人脸区域的多个指标并确定多个指标的权重，否则，重新执行所述在当前逆光补偿参数下采集人脸图像，并在采集到的人脸图像上定位出人脸区域。Determine whether the face image is stable according to the changes of the collected multiple frames of face images, if so, execute the calculation of multiple indicators of the face area and determine the weights of the multiple indicators, otherwise, re-execute the current backlight compensation parameter Then, the face image is collected, and the face area is located on the collected face image.

进一步的，所述根据采集到的多帧人脸图像的变化判断人脸图像是否稳定包括：Further, judging whether the face image is stable according to the changes of the collected multi-frame face images includes:

将人脸图像减去人脸区域，得到环境区域；Subtract the face area from the face image to get the environment area;

分别计算两帧人脸图像的人脸区域和环境区域的灰度方差变化值；Calculate the change value of the grayscale variance of the face area and the environment area of the two frames of face images respectively;

若两帧人脸图像的人脸区域和环境区域的灰度方差变化值分别小于设定阈值，则人脸图像稳定。If the variation values of the grayscale variance of the face region and the environment region of the two frames of face images are respectively smaller than the set thresholds, the face image is stable.

进一步的，所述分别计算两帧人脸图像的人脸区域和环境区域的灰度方差变化值包括：Further, the calculation of the grayscale variance change values of the face area and the environment area of the two frames of face images respectively includes:

计算相邻两帧人脸图像的人脸区域/环境区域的灰度方差；Calculate the grayscale variance of the face area/environment area of two adjacent frames of face images;

将相邻两帧人脸图像的人脸区域/环境区域的灰度方差相减后求绝对值，得到人脸区域/环境区域的灰度方差变化值；Calculate the absolute value after subtracting the grayscale variance of the face area/environmental area of two adjacent frames of face images, and obtain the change value of the grayscale variance of the face area/environmental area;

按照多帧人脸图像的采集时间顺序每隔一定帧数计算一次人脸区域/环境区域的灰度方差变化值，得到按时间顺序的人脸区域/环境区域的多个灰度方差变化值；According to the acquisition time sequence of the multi-frame face images, calculate the grayscale variance change values of the face area/environmental area every certain number of frames, and obtain multiple grayscale variance change values of the face area/environmental area in chronological order;

若人脸区域的多个灰度方差变化值均小于设定的人脸区域灰度方差变化阈值，且时间顺序靠后的人脸区域的灰度方差变化值小于时间顺序靠前的人脸区域的灰度方差变化值；同时，环境区域的多个灰度方差变化值均小于设定的环境区域灰度方差变化阈值，且时间顺序靠后的环境区域的灰度方差变化值小于时间顺序靠前的环境区域的灰度方差变化值；则人脸图像稳定。If the multiple gray variance change values of the face region are all smaller than the set threshold value of the gray variance variance of the face region, and the gray variance change value of the face region in the later time sequence is smaller than that of the face region earlier in the time sequence At the same time, the multiple gray variance change values of the environmental area are all smaller than the set threshold value of the gray variance change of the environmental area, and the gray variance change value of the environmental area in the later time sequence is smaller than that in the later time sequence. The gray variance change value of the previous environment area; then the face image is stable.

进一步的，所述人脸区域灰度指标包括人脸区域的亮部灰度占比、人脸T区的灰度方差、人脸T区的灰度均值、人脸区域的灰度动态范围、人脸区域的二维熵，所述人脸区域频率域指标包括人脸区域的FFT高频占比，所述多个指标包括人脸区域的亮部灰度占比、人脸T区的灰度方差、人脸区域的灰度动态范围、人脸区域的二维熵、人脸区域的FFT高频占比中的至少一个以及人脸T区的灰度均值。Further, the grayscale index of the face region includes the proportion of the grayscale of the bright part of the face region, the grayscale variance of the face T region, the grayscale mean of the face T region, the grayscale dynamic range of the face region, The two-dimensional entropy of the face region, the frequency domain index of the face region includes the FFT high frequency ratio of the face region, and the multiple indicators include the grayscale ratio of the bright parts of the face region, the grayscale of the face T region. At least one of the degree variance, the gray dynamic range of the face region, the two-dimensional entropy of the face region, the high frequency ratio of the FFT of the face region, and the gray mean value of the T-zone of the face.

进一步的，调整当前逆光补偿参数的方法为：Further, the method for adjusting the current backlight compensation parameters is:

根据调整参数与预设的阈值范围的上限或下限的差值确定调整数值，使用确定出的调整数值一次调整逆光补偿参数；Determine the adjustment value according to the difference between the adjustment parameter and the upper limit or lower limit of the preset threshold range, and use the determined adjustment value to adjust the backlight compensation parameter at one time;

或者，根据调整参数小于预设的阈值范围的下限或大于预设的阈值范围的上限的情况将当前逆光补偿参数增加或减小固定数值，然后返回最开始的步骤，重复调整逆光补偿参数，直至调整参数在预设的阈值范围内。Alternatively, increase or decrease the current backlight compensation parameter by a fixed value according to the condition that the adjustment parameter is smaller than the lower limit of the preset threshold range or larger than the upper limit of the preset threshold range, and then returns to the initial step, and repeatedly adjusts the backlight compensation parameter until Adjust the parameters within the preset threshold range.

第二方面，本发明提供一种人脸识别逆光补偿装置，所述装置包括：In a second aspect, the present invention provides a face recognition backlight compensation device, the device comprising:

采集模块，用于在当前逆光补偿参数下采集人脸图像，并在采集到的人脸图像上定位出人脸区域；The acquisition module is used to collect the face image under the current backlight compensation parameters, and locate the face area on the collected face image;

指标计算模块，用于计算人脸区域的多个指标并确定多个指标的权重，所述多个指标包括人脸区域灰度指标和/或人脸区域频率域指标；an index calculation module for calculating multiple indices of the face region and determining the weights of the multiple indices, the multiple indices including the face region grayscale index and/or the face region frequency domain index;

调整参数计算模块，用于根据所述人脸区域的多个指标以及多个指标的权重计算调整参数；an adjustment parameter calculation module, configured to calculate adjustment parameters according to the multiple indicators of the face region and the weights of the multiple indicators;

调整模块，用于判断调整参数是否在预设的阈值范围内，若是，不调整当前逆光补偿参数，否则，调整当前逆光补偿参数。The adjustment module is used for judging whether the adjustment parameters are within the preset threshold range, if so, do not adjust the current backlight compensation parameters, otherwise, adjust the current backlight compensation parameters.

进一步的，在当前逆光补偿参数下采集到的人脸图像为多帧，所述装置还包括：Further, the face images collected under the current backlight compensation parameters are multiple frames, and the device further includes:

稳定判断模块，用于根据采集到的多帧人脸图像的变化判断人脸图像是否稳定，若是，执行所述指标计算模块，否则，重新执行所述采集模块；a stability judging module, configured to judge whether the face image is stable according to the changes of the collected multi-frame face images, if yes, execute the index calculation module, otherwise, execute the collection module again;

所述稳定判断模块包括：The stability judgment module includes:

环境区域获取单元，用于将人脸图像减去人脸区域，得到环境区域；The environmental area acquisition unit is used to subtract the face area from the face image to obtain the environmental area;

灰度方差变化值计算单元，用于分别计算两帧人脸图像的人脸区域和环境区域的灰度方差变化值；A grayscale variance change value calculation unit, used to calculate the grayscale variance change values of the face area and the environment area of the two frames of face images respectively;

其中，若两帧人脸图像的人脸区域和环境区域的灰度方差变化值分别小于设定阈值，则人脸图像稳定。Wherein, if the change values of the grayscale variance of the face region and the environment region of the two frames of face images are respectively smaller than the set threshold, the face image is stable.

第三方面，本发明提供一种用于人脸识别逆光补偿的计算机可读存储介质，包括处理器及用于存储处理器可执行指令的存储器，所述指令被所述处理器执行时实现包括前述的第一方面所述人脸识别逆光补偿方法的步骤。In a third aspect, the present invention provides a computer-readable storage medium for face recognition backlight compensation, comprising a processor and a memory for storing processor-executable instructions, where the instructions, when executed by the processor, include: The steps of the face recognition backlight compensation method described in the first aspect.

第四方面，本发明提供一种用于人脸识别逆光补偿的设备，包括至少一个处理器以及存储计算机可执行指令的存储器，所述处理器执行所述指令时实现前述的第一方面所述人脸识别逆光补偿方法的步骤。In a fourth aspect, the present invention provides a device for face recognition backlight compensation, comprising at least one processor and a memory for storing computer-executable instructions, the processor implementing the above-mentioned first aspect when executing the instructions. The steps of the face recognition backlight compensation method.

本发明具有以下有益效果：The present invention has the following beneficial effects:

本发明采用人脸区域灰度指标和/或人脸区域频率域指标及其权重计算得到能够很好地反映人脸区域清晰度的调整参数，并根据调整参数与设定的阈值范围的关系决定是否调整当前逆光补偿参数，以及如何调整当前逆光补偿参数，得到合适的逆光补偿参数，在该逆光补偿参数的补光下采集到清晰的人脸图像。本发明适用于室内暗光、室内强逆光、室外强逆光、夜晚室外环境下的自动补光，能够自适应调节逆光补偿参数，解决暗光及逆光环境下人脸图像过曝或过暗从而使得人脸面部信息缺失的问题。The present invention adopts the face area grayscale index and/or the face area frequency domain index and its weight to calculate the adjustment parameter that can well reflect the clarity of the face area, and determines the adjustment parameter according to the relationship between the adjustment parameter and the set threshold range Whether to adjust the current backlight compensation parameters, and how to adjust the current backlight compensation parameters to obtain appropriate backlight compensation parameters, and collect clear face images under the fill light of the backlight compensation parameters. The invention is suitable for automatic fill light in indoor dark light, indoor strong backlight, outdoor strong backlight, and outdoor environment at night, and can adaptively adjust the backlight compensation parameters, so as to solve the problem of overexposure or too dark of the face image in the dark light and backlight environment, so that the The problem of missing face information.

附图说明Description of drawings

图1为本发明的人脸识别逆光补偿方法流程图；Fig. 1 is the flow chart of the face recognition backlight compensation method of the present invention;

图2为68特征点定位的方法示意图；Fig. 2 is the method schematic diagram of 68 feature point positioning;

图3为本发明的人脸识别逆光补偿装置示意图。FIG. 3 is a schematic diagram of a face recognition backlight compensation device of the present invention.

具体实施方式Detailed ways

为使本发明要解决的技术问题、技术方案和优点更加清楚，下面将结合附图及具体实施例对本发明的技术方案进行清楚、完整地描述。显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。通常在此处附图中描述和示出的本发明实施例的组件可以以各种不同的配置来布置和设计。因此，以下对在附图中提供的本发明的实施例的详细描述并非旨在限制要求保护的本发明的范围，而是仅仅表示本发明的选定实施例。基于本发明的实施例，本领域技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例，都属于本发明保护的范围。In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, but not all, embodiments of the present invention. The components of the embodiments of the invention generally described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations. Thus, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but is merely representative of selected embodiments of the invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present invention.

实施例1：Example 1:

本发明实施例提供了一种人脸识别逆光补偿方法，用于室内暗光、室内强逆光、室外强逆光、夜晚室外环境下自动调整逆光补偿参数，从而能够清晰成像。如图1所示，该方法包括：The embodiment of the present invention provides a face recognition backlight compensation method, which is used for automatic adjustment of backlight compensation parameters in indoor dark light, indoor strong backlight, outdoor strong backlight, and outdoor environment at night, thereby enabling clear imaging. As shown in Figure 1, the method includes:

步骤S100：在当前逆光补偿参数下采集人脸图像，并在采集到的人脸图像上定位出人脸区域。Step S100: Collect a face image under the current backlight compensation parameter, and locate a face area on the collected face image.

在本步骤中，最初的时候使用当前逆光补偿参数进行补光，并在当前逆光补偿参数下采集人脸图像。本发明实施例的人脸图像仅仅是指人脸识别装置的摄像机或摄像头采集到的图像，并不是指包括人脸的图像，人脸图像可以包括人脸，也可以不包括人脸，例如用户人脸在摄像头之外时，采集到的图像不包括人脸，这类的图像在本发明中也被称为人脸图像。In this step, initially, the current backlight compensation parameters are used to fill in the light, and a face image is collected under the current backlight compensation parameters. The face image in the embodiment of the present invention only refers to the image collected by the camera or the camera of the face recognition device, and does not refer to the image including the face. The face image may include the face, or may not include the face. For example, a user When the human face is outside the camera, the collected image does not include the human face, and such images are also referred to as human face images in the present invention.

本步骤中，优选通过人脸检测算法检测人脸图像里是否存在人脸，并定位出人脸区域。In this step, it is preferable to use a face detection algorithm to detect whether there is a human face in the face image, and to locate the face region.

步骤S200：计算人脸区域的多个指标并确定多个指标的权重，多个指标包括人脸区域灰度指标和/或人脸区域频率域指标。Step S200: Calculate multiple indices of the face region and determine the weights of the multiple indices, where the multiple indices include a face region grayscale index and/or a face region frequency domain index.

灰度指标和人脸区域频率域指标能够反映人脸区域的清晰度，本发明实施例可以根据需要选择合适的指标；指标权重可以按照指标的数量，各指标对总体的贡献度以及各指标对补光的灵敏度等进行设置。The grayscale index and the frequency domain index of the face region can reflect the clarity of the face region, and in this embodiment of the present invention, an appropriate index can be selected according to needs; Set the sensitivity of fill light, etc.

步骤300：根据人脸区域的多个指标以及多个指标的权重计算调整参数。Step 300: Calculate the adjustment parameters according to the multiple indicators of the face region and the weights of the multiple indicators.

在本步骤中，在获得人脸区域的多个指标以及每个指标对应的权重之后，可以根据多个指标中的全部指标及其对应的权重计算调整参数，也可以根据多个指标中的部分指标及其对应的权重计算调整参数。调整参数将各个指标按权重进行综合，能够很好地反映人脸区域的清晰度。In this step, after obtaining the multiple indicators of the face area and the corresponding weight of each indicator, the adjustment parameters may be calculated according to all the indicators in the multiple indicators and their corresponding weights, or the adjustment parameters may be calculated according to some of the multiple indicators. Indicators and their corresponding weight calculation adjustment parameters. Adjusting the parameters integrates each index according to the weight, which can well reflect the clarity of the face area.

步骤400：判断调整参数是否在预设的阈值范围内，若是，则人脸图像的清晰度满足要求，不调整当前逆光补偿参数，输出采集到的人脸图像即可，否则，人脸图像的清晰度不满足要求，调整当前逆光补偿参数，使用调整后的逆光补偿参数进行补光。Step 400: Judging whether the adjustment parameters are within the preset threshold range, if so, the clarity of the face image meets the requirements, do not adjust the current backlight compensation parameters, and output the collected face image, otherwise, the face image is not adjusted. If the clarity does not meet the requirements, adjust the current backlight compensation parameters, and use the adjusted backlight compensation parameters to fill in the light.

当前逆光补偿参数的调整根据调整参数与预设的阈值范围的上下限的关系进行相应的增加或减小，具体是增加还是减小，以及增加减小的数值则与指标的选择以及预设的阈值范围的设定相关，调整的结果是使得人脸图像更清晰，调整参数向预设的阈值范围靠近。The adjustment of the current backlight compensation parameter is correspondingly increased or decreased according to the relationship between the adjustment parameter and the upper and lower limits of the preset threshold range. The setting of the threshold range is related. The result of the adjustment is to make the face image clearer, and the adjustment parameters are closer to the preset threshold range.

当步骤S100中检测到人脸图像中存在人脸时，还可以进一步判断采集到的人脸图像是否稳定，判断人脸图像是否稳定需要用到多幅人脸图像，本发明实施例优选通过视频流获取多帧人脸图像，此时，步骤S100中，在当前逆光补偿参数下采集到的人脸图像为多帧，判断人脸图像是否稳定的方法包括：When it is detected that there is a human face in the face image in step S100, it can be further judged whether the collected face image is stable, and multiple face images need to be used to judge whether the face image is stable. Flow to obtain multiple frames of face images, at this time, in step S100, the face images collected under the current backlight compensation parameters are multiple frames, and the method for judging whether the face images is stable includes:

步骤S110：根据采集到的多帧人脸图像的变化判断人脸图像是否稳定，若是，执行步骤S200，否则，重新执行步骤S100。Step S110: Determine whether the face image is stable according to the changes of the collected multiple frames of the face image, if yes, go to step S200, otherwise, go to step S100 again.

本发明的采集帧速优选为30帧/S，采集到的多帧人脸图像的分辨率优选为640*480。此时，优选采用68特征点定位的方法进行人脸检测，68特征点定位的人脸检测方法执行效率快，丝毫不影响设定的采集帧速30帧/s，能够避免漏检和误检，从而提高人脸检测的执行效率，当然，也可以采用其他人脸检测方法。The acquisition frame rate of the present invention is preferably 30 frames/S, and the resolution of the collected multi-frame face images is preferably 640*480. At this time, it is preferable to use the method of 68 feature point positioning for face detection. The face detection method of 68 feature point positioning is fast in execution and does not affect the set acquisition frame rate of 30 frames/s, which can avoid missed detection and false detection. , so as to improve the execution efficiency of face detection. Of course, other face detection methods can also be used.

标定人脸68点示意图如图2所示，依次连线点1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17-27-26-25-24-23-22-21-20-19-18-1所围成的闭合区域定义为人脸区域。其中，当检测到多张人脸的时候，取点1到点17的距离作为人脸宽度，取宽度最大者为目标人脸对象。The schematic diagram for calibrating 68 points of the face is shown in Figure 2, connecting the points 1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17- The closed area enclosed by 27-26-25-24-23-22-21-20-19-18-1 is defined as the face area. Wherein, when multiple faces are detected, the distance from point 1 to point 17 is taken as the face width, and the one with the largest width is taken as the target face object.

人脸图像稳定是指采集到的多帧人脸图像相似，其差别不大，说明在采集这一系列人脸图像时，用户人脸、摄像头以及环境光照等处于稳定的状态，采集到的人脸图像的质量以及一致性较好。本发明实施例通过采集到的多帧人脸图像的变化判断人脸图像是否稳定。Face image stabilization means that the collected multiple frames of face images are similar, and the difference is not large, which means that when collecting this series of face images, the user's face, camera, and ambient light are in a stable state. The quality and consistency of the face images are good. In the embodiment of the present invention, it is determined whether the face image is stable or not according to the changes of the collected multiple frames of the face image.

本发明中，根据采集到的多帧人脸图像的变化判断人脸图像是否稳定的方法可以有多种，可以选取人脸图像的多种特征指标来反映多帧人脸图像的变化，这里给出一个示例，包括：In the present invention, there can be various methods for judging whether the face image is stable according to the changes of the collected multi-frame face images, and various characteristic indicators of the face image can be selected to reflect the changes of the multi-frame face images. Give an example, including:

步骤S111：将人脸图像减去人脸区域，得到环境区域。Step S111 : subtract the face area from the face image to obtain the environment area.

在本步骤中，人脸区域指的是整张人脸图像中的人脸部分，环境区域为整张人脸图像中去掉人脸区域外的区域。In this step, the face area refers to the face part in the whole face image, and the environment area is the area outside the face area in the whole face image.

步骤S112：分别计算两帧人脸图像的人脸区域和环境区域的灰度方差变化值。Step S112: Calculate the grayscale variance change values of the face region and the environment region of the two frames of face images respectively.

在本步骤中，两帧图像的灰度方差变化值反应的是两帧图像亮度的变化值，本发明实施例根据图像亮度变化值判断图像是否稳定。两帧人脸图像是预设帧，优选是一秒内采集的人脸图像中的两帧，更优选的是相邻的两帧。In this step, the change value of the grayscale variance of the two frames of images reflects the change value of the brightness of the two frames of images, and the embodiment of the present invention determines whether the image is stable according to the change value of the image brightness. The two frames of face images are preset frames, preferably two frames in the face images collected within one second, and more preferably two adjacent frames.

其中，若两帧人脸图像的人脸区域和环境区域的灰度方差变化值分别小于设定阈值，则人脸图像稳定。即两帧人脸图像的人脸区域和环境区域亮度变化小于一定程度即认为人脸图像稳定。Wherein, if the change values of the grayscale variance of the face region and the environment region of the two frames of face images are respectively smaller than the set threshold, the face image is stable. That is, the face image is considered to be stable if the brightness changes of the face area and the environment area of the two frames of face images are less than a certain degree.

具体的，分别计算两帧人脸图像的人脸区域和环境区域的灰度方差变化值包括：Specifically, calculating the gray variance change values of the face region and the environment region of the two frames of face images respectively includes:

步骤S1121：计算相邻两帧人脸图像的人脸区域(或环境区域)的灰度方差。Step S1121: Calculate the grayscale variance of the face area (or the environment area) of two adjacent frames of face images.

灰度方差反映一幅图像的清晰度，如果一幅图像的灰度方差小，那么该图像看起来灰蒙蒙的，不清晰；如果一幅图像的灰度方差大，那么该图像看起来对比度明显，清晰。The grayscale variance reflects the sharpness of an image. If the grayscale variance of an image is small, the image looks gray and unclear; if the grayscale variance of an image is large, the image looks obvious contrast. , clear.

本发明实施例中，灰度方差可以是现有技术常规的图像灰度方差，即每个像素点的灰度值减去图像平均灰度值的平方和除以总的像素个数，也可以是本发明定义的灰度方差，参见后文的定义。In the embodiment of the present invention, the grayscale variance may be the conventional image grayscale variance in the prior art, that is, the grayscale value of each pixel point minus the square sum of the average grayscale value of the image divided by the total number of pixels, or it may be is the grayscale variance defined in the present invention, see the definition below.

步骤S1122：将相邻两帧人脸图像的人脸区域(或环境区域)的灰度方差相减后求绝对值，即可得到人脸区域(或环境区域)的灰度方差变化值。Step S1122: Subtract the grayscale variances of the face regions (or environmental regions) of two adjacent frames of face images and then obtain the absolute value to obtain the grayscale variance change values of the face regions (or environmental regions).

本发明可以通过一个灰度方差变化值判断人脸图像是否稳定，此时只要判断相邻两帧人脸图像的人脸区域和环境区域的灰度方差变化值是否分别小于设定阈值即可。The present invention can judge whether the face image is stable through one grayscale variance change value. In this case, it is only necessary to judge whether the grayscale variance change values of the face area and the environment area of two adjacent frames of face images are respectively smaller than the set threshold.

本发明还可以通过一系列的灰度方差变化值判断人脸图像是否稳定，此时，本发明实施例还包括：The present invention can also judge whether the face image is stable through a series of grayscale variance change values. At this time, the embodiment of the present invention further includes:

步骤S1123：按照多帧人脸图像的采集时间顺序每隔一定帧数计算一次人脸区域(或环境区域)的灰度方差变化值，得到按时间顺序的人脸区域(或环境区域)的多个灰度方差变化值。Step S1123: Calculate the change value of the grayscale variance of the face area (or the environment area) at intervals of a certain number of frames according to the collection time sequence of the multi-frame face images, and obtain the chronological order of the face area (or the environment area). A grayscale variance change value.

其中，若人脸区域的多个灰度方差变化值均小于设定的人脸区域灰度方差变化阈值，且时间顺序靠后的人脸区域的灰度方差变化值小于时间顺序靠前的人脸区域的灰度方差变化值；同时，环境区域的多个灰度方差变化值均小于设定的环境区域灰度方差变化阈值，且时间顺序靠后的环境区域的灰度方差变化值小于时间顺序靠前的环境区域的灰度方差变化值；则人脸图像稳定。Among them, if the multiple grayscale variance change values of the face region are all smaller than the set grayscale variance change threshold of the face region, and the grayscale variance change value of the face region in the later time sequence is smaller than that of the person in the earlier time sequence The gray variance change value of the face area; at the same time, the multiple gray variance change values of the environmental area are all smaller than the set environmental area gray variance change threshold, and the gray variance change value of the environmental area in the later time sequence is smaller than the time The change value of the grayscale variance of the environmental area at the front of the order; the face image is stable.

本发明通过一系列的按时间顺序的灰度方差变化值来判断人脸图像是否稳定，更能反映出人脸图像趋于稳定的过程，判断结果更准确。The invention judges whether the face image is stable through a series of time-sequential grayscale variance change values, which can better reflect the process of the face image tending to be stable, and the judgment result is more accurate.

本发明的多个指标可以根据需要进行设定，例如：人脸区域灰度指标可以包括人脸区域的亮部灰度占比、人脸T区的灰度方差、人脸T区的灰度均值、人脸区域的灰度动态范围、人脸区域的二维熵，人脸区域频率域指标可以包括人脸区域的FFT高频占比。Multiple indicators of the present invention can be set as required, for example: the gray scale indicators of the face area may include the gray scale ratio of the bright part of the face area, the gray scale variance of the face T area, and the gray scale of the face T area. The mean value, the gray dynamic range of the face area, the two-dimensional entropy of the face area, and the frequency domain index of the face area may include the high frequency ratio of the FFT of the face area.

多个指标优选必须包括人脸T区的灰度均值，以及还可以包括人脸区域的亮部灰度占比、人脸T区的灰度方差、人脸区域的灰度动态范围、人脸区域的二维熵、人脸区域的FFT高频占比中的至少一个。The multiple indicators preferably must include the gray mean value of the face T area, and can also include the gray scale ratio of the bright part of the face area, the gray scale variance of the face T area, the gray dynamic range of the face area, and the gray scale of the face area. At least one of the two-dimensional entropy of the region and the high frequency ratio of the FFT of the face region.

本发明中，各个指标的计算方法如下：In the present invention, the calculation method of each index is as follows:

人脸区域的亮部灰度占比通过如下方法计算得到：The gray-scale proportion of the bright part of the face area is calculated by the following method:

步骤S210：图像每个像素的灰度值的范围为0～255，将人脸区域按灰度亮阶分区为三种亮阶区域，其中：Step S210 : the gray value of each pixel of the image ranges from 0 to 255, and the face area is divided into three types of bright-level areas according to the gray-level brightness level, wherein:

第一亮阶区域：0≤gray[x,y]≤150；The first bright level area: 0≤gray[x,y]≤150;

第二亮阶区域：151≤gray[x,y]≤220；The second bright-level area: 151≤gray[x,y]≤220;

第三亮阶区域：221≤gray[x,y]≤255；The third bright-level area: 221≤gray[x,y]≤255;

gray[x,y]为人脸区域第x行，第y列的像素的灰度值；gray[x,y] is the gray value of the pixel in the xth row and the yth column of the face area;

步骤S220：计算三种亮阶区域占人脸区域的占比brArRoF、brArRoS、brArRoT，本发明实施例优选将第二亮阶区域的占比brArRoS作为人脸区域的亮部灰度占比。Step S220: Calculate the proportions of the three bright-level regions in the face region, brArRoF, brArRoS, and brArRoT. In the embodiment of the present invention, the proportion of the second bright-level region, brArRoS, is preferably used as the bright gray proportion of the face region.

本发明实施例还包括：Embodiments of the present invention also include:

步骤S230：根据灰度方差的定义公式计算人脸区域的灰度方差、背景区域的灰度方差和人脸T区的灰度方差，本发明实施例图像的灰度方差的定义公式优选如下：Step S230: Calculate the grayscale variance of the face area, the grayscale variance of the background area, and the grayscale variance of the face T area according to the definition formula of the grayscale variance. The definition formula of the grayscale variance of the image in the embodiment of the present invention is preferably as follows:

variGy＝(gray[x1,y1]-AveGy)×(gray[x1,y1]-AveGy)+variGy=(gray[x1,y1]-AveGy)×(gray[x1,y1]-AveGy)+

(gray[x1+1,y1]-AveGy)×(gray[x1+2,y1]-AveGy)+(gray[x1+1,y1]-AveGy)×(gray[x1+2,y1]-AveGy)+

...+...+

(gray[x2,y1]-AveGy)×(gray[x2,y1]-AveGy)+(gray[x2,y1]-AveGy)×(gray[x2,y1]-AveGy)+

(gray[x1,y1+1]-AveGy)×(gray[x1,y1+1]-AveGy)+(gray[x1,y1+1]-AveGy)×(gray[x1,y1+1]-AveGy)+

(gray[x1+1,y1+1]-AveGy)×(gray[x1+2,y1+1]-AveGy)+(gray[x1+1,y1+1]-AveGy)×(gray[x1+2,y1+1]-AveGy)+

...+...+

(gray[x2,y1+1]-AveGy)×(gray[x2,y1+1]-AveGy)+(gray[x2,y1+1]-AveGy)×(gray[x2,y1+1]-AveGy)+

......+......+

(gray[x1,y2]-AveGy)×(gray[x1,y2]-AveGy)+(gray[x1,y2]-AveGy)×(gray[x1,y2]-AveGy)+

(gray[x1+1,y2]-AveGy)×(gray[x1+2,y2]-AveGy)+(gray[x1+1,y2]-AveGy)×(gray[x1+2,y2]-AveGy)+

...+...+

(gray[x2,y2]-AveGy)×(gray[x2,y2]-AveGy)(gray[x2,y2]-AveGy)×(gray[x2,y2]-AveGy)

(x1，y1)为图像的左下角坐标，(x2，y2)为图像的右上角坐标，AveGy为图像的灰度均值，gray[x,y]为图像坐标(x,y)像素的灰度值，x＝x1,x1+1,x1+2,……,x2，y＝y1,y1+1,y1+2,……,y2。(x1, y1) is the coordinate of the lower left corner of the image, (x2, y2) is the coordinate of the upper right corner of the image, AveGy is the grayscale mean of the image, and gray[x,y] is the grayscale of the pixel at the image coordinate (x,y) Value, x=x1, x1+1, x1+2,...,x2, y=y1, y1+1, y1+2,...,y2.

以人脸图像大小为640*480分辨率，每秒采集30帧为例来说明人脸区域的灰度方差以及灰度方差变化值的计算过程：Taking the face image size as 640*480 resolution and collecting 30 frames per second as an example to illustrate the calculation process of the grayscale variance of the face area and the change value of the grayscale variance:

首先，计算人脸区域的灰度均值AveGyF，将人脸区域的各个像素点的灰度值加权平均即可。First, calculate the average gray value AveGyF of the face area, and then weight the average gray value of each pixel in the face area.

然后，根据灰度方差的定义公式计算前后相邻两帧人脸图像的人脸区域的灰度方差variGyF[1]和variGyF[2]。Then, according to the definition formula of the gray variance, the gray variances variGyF[1] and variGyF[2] of the face regions of the two adjacent frames of face images are calculated.

根据variGyF[1]和variGyF[2]计算人脸区域的灰度方差变化值，diffVariGyF，diffVariGyF＝∣variGyF[1]-variGyF[2]∣。Calculate the grayscale variance change value of the face area according to variGyF[1] and variGyF[2], diffVariGyF, diffVariGyF=∣variGyF[1]-variGyF[2]∣.

如果需要时间先后顺序的多个灰度方差变化值，则还进行如下操作：If multiple grayscale variance change values in chronological order are required, the following operations are also performed:

设备的帧速为每秒30帧，算得当前秒内，每隔10帧的灰度方差变化值如下：The frame rate of the device is 30 frames per second. Calculated in the current second, the grayscale variance change value of every 10 frames is as follows:

diffVariGyF[1]＝∣variGyF[n+1]-variGyF[n+0]∣；diffVariGyF[1]=∣variGyF[n+1]-variGyF[n+0]∣;

diffVariGyF[2]＝∣variGyF[n+11]-variGyF[n+10]∣；diffVariGyF[2]=∣variGyF[n+11]-variGyF[n+10]∣;

diffVariGyF[3]＝∣variGyF[n+21]-variGyF[n+20]∣；diffVariGyF[3]=∣variGyF[n+21]-variGyF[n+20]∣;

n为常数。n is a constant.

以相同的方法计算得到按时间顺序的环境区域的多个灰度方差变化值diffVariGyE[1],diffVariGyE[2],diffVariGyE[3]；Calculate in the same way to obtain multiple grayscale variance change values diffVariGyE[1], diffVariGyE[2], diffVariGyE[3] of the environmental area in time sequence;

若同时满足以下两个条件，则人脸图像稳定：If the following two conditions are met at the same time, the face image is stable:

diffVariGyF[3]＜diffVariGyF[2]＜diffVariGyF[1]＜diffMaxF；diffVariGyF[3]<diffVariGyF[2]<diffVariGyF[1]<diffMaxF;

diffVariGyE[3]＜diffVariGyE[2]＜diffVariGyE[1]＜diffMaxE；diffVariGyE[3]<diffVariGyE[2]<diffVariGyE[1]<diffMaxE;

diffMaxF为设定的人脸区域的灰度方差变化阈值，diffMaxE为设定的环境区域的灰度方差变化阈值。两个阈值是视调试情况来确定的经验参数。diffMaxF is the set threshold value of the gray variance of the face area, and diffMaxE is the set threshold of the gray variance of the environment area. The two thresholds are empirical parameters determined depending on the debugging situation.

人脸T区是指人脸的额头区域和鼻子区域组成的T形区域，在采集到的人脸图像上定位出人脸区域也就能标定出人脸T区，例如68特征点定位出的人脸可以很明显的得到人脸T区，如图2所示。The face T area refers to the T-shaped area composed of the forehead area and the nose area of the human face. If the face area is located on the collected face image, the face T area can also be demarcated. For example, the 68 feature points are located. The face can obviously get the T-zone of the face, as shown in Figure 2.

当判断采集的人脸图像稳定时，进行人脸标定，并标定出人脸T区，根据前述的图像的灰度方差的定义公式即可计算出人脸T区的灰度方差variGyT。When it is judged that the collected face image is stable, the face calibration is performed, and the face T area is calibrated. According to the above-mentioned definition formula of the gray scale variance of the image, the gray scale variance variGyT of the face T area can be calculated.

将人脸T区的各个像素点的灰度值加权平均即可得到人脸T区的灰度均值AveGyT。The average gray value AveGyT of the face T area can be obtained by weighted average of the gray value of each pixel in the face T area.

人脸区域的灰度动态范围通过如下方法计算得到：The gray dynamic range of the face region is calculated by the following method:

步骤S240：统计得到人脸区域的有效最小灰度值和有效最大灰度值，其中：灰度值的像素个数占总像素个数的占比达到设定比值的最小灰度值即为有效最小灰度值，灰度值的像素个数占总像素个数的占比达到设定比值的最大灰度值即为有效最大灰度值。Step S240: Statistically obtain the effective minimum grayscale value and the effective maximum grayscale value of the face region, wherein: the minimum grayscale value whose ratio of the number of grayscale pixels to the total number of pixels reaches the set ratio is considered valid. The minimum grayscale value, the ratio of the number of grayscale pixels to the total number of pixels, and the maximum grayscale value that reaches the set ratio is the effective maximum grayscale value.

本步骤对人脸区域的不同灰度值进行统计，取得有效最小灰度值minGy和有效最大灰度值maxGy，In this step, the different gray values of the face area are counted, and the effective minimum gray value minGy and the effective maximum gray value maxGy are obtained,

有效最小灰度值是指达到此灰度值minGy的像素个数在人脸区域占一定比例，本实施例中取的比例值为1％；有效最大灰度值同理。dyRgGy的理论范围在0～255。The effective minimum gray value means that the number of pixels reaching this gray value minGy occupies a certain proportion in the face area, and the proportion taken in this embodiment is 1%; the same is true for the effective maximum gray value. The theoretical range of dyRgGy is 0-255.

步骤S250：将有效最大灰度值与有效最小灰度值相减即得到人脸区域的灰度动态范围dyRgGy，公式如下：Step S250: subtracting the effective maximum gray value and the effective minimum gray value to obtain the gray dynamic range dyRgGy of the face region, the formula is as follows:

dyRgGy＝maxGy-minGy。dyRgGy=maxGy-minGy.

人脸图像的二维熵通过如下方法计算得到：The two-dimensional entropy of the face image is calculated by the following method:

步骤S260：首先进行灰度直方图的统计，即统计人脸区域的灰度级的分布情况，记录下来0～255这个区间内每一灰度级的像素出现的个数，nH[i],i＝0,1,2,……,255，以向量的形式记为nH，按式hG[i]＝nH[i]/nSgy算得向量hG，向量hG长度为255，向量hG的元素hG[i]代表灰度级为i的像素在人脸区域中出现的频率(其中nSgy为人脸区域总像素数)，人脸区域的一维熵imgEpyO的计算公式如下：Step S260: First, perform the statistics of the grayscale histogram, that is, the distribution of the grayscale levels in the face region, and record the number of pixels appearing in each grayscale level in the range of 0 to 255, nH[i], i=0, 1, 2,..., 255, denoted as nH in the form of a vector, the vector hG is calculated according to the formula hG[i]=nH[i]/nSgy, the length of the vector hG is 255, and the element of the vector hG hG[ i] represents the frequency of the pixel whose gray level is i in the face area (where nSgy is the total number of pixels in the face area). The one-dimensional entropy imgEpyO of the face area is calculated as follows:

步骤S270：记i为中心像素的灰度级(范围0～255)，j为邻域像素的灰度级(范围0～255)，定义一个特征二元组hHT[i，j]记录人脸区域中中心灰度级为i，邻域像素灰度级为j的像素在人脸区域出现的次数，N为图像尺度，按式hGT[i，j]＝hHT[i，j]/N²算得hGT[i，j]为人脸区域中中心灰度级为i，邻域像素灰度级为j的像素在人脸区域出现的频率；则图像二维熵imgEpyT的计算公式如下，Step S270: Denote i as the gray level of the center pixel (range 0-255), j as the gray level of the neighboring pixels (range 0-255), and define a feature two-tuple hHT[i, j] to record the face The number of times that the gray level of the center of the area is i, and the gray level of the neighboring pixels is j in the face area, N is the image scale, according to the formula hGT[i, j]=hHT[i, j]/N² It is calculated that hGT[i, j] is the frequency of the pixels whose central gray level is i and the neighborhood pixel gray level is j in the face region in the face region; then the calculation formula of the two-dimensional image entropy imgEpyT is as follows:

人脸区域的FFT高频占比通过如下方法计算得到：The FFT high frequency proportion of the face area is calculated by the following method:

步骤S280：对人脸区域进行二维离散傅里叶变换(Fast FourierTransformation,FFT)，将人脸区域从空间域转换成频率域；二维离散傅立叶变换公式如下：Step S280: Perform a two-dimensional discrete Fourier transform (Fast Fourier Transform, FFT) on the face area, and convert the face area from the spatial domain to the frequency domain; the two-dimensional discrete Fourier transform formula is as follows:

其中，i为虚数单位，u＝0,1,...,M-1,v＝0,1,...N-1；f(m,n)为原图像的横坐标m，纵坐标n的像素的灰度值，F(u,v)为FFT后图像的横坐标u,纵坐标v的像素的灰度值，M为图像区域的宽度，N为图像区域的高度。Among them, i is the imaginary unit, u=0,1,...,M-1,v=0,1,...N-1; f(m,n) is the abscissa m of the original image, and the ordinate The gray value of the pixel of n, F(u, v) is the gray value of the abscissa u of the image after FFT, and the gray value of the pixel of the ordinate v, M is the width of the image area, and N is the height of the image area.

步骤S290：统计频率域的高频分量占比，即为人脸区域的FFT高频占比fftRo。Step S290: Count the proportion of high frequency components in the frequency domain, that is, the FFT high frequency proportion fftRo of the face region.

本发明在调试中测试大量的人脸图像，发现成像清晰的图像，其高频分量占比比较高，得到fftRo后，设定合适的区间fftRoMin～fftRoMax，取落入此合理区间的fftRo作为后续步骤的一个评价指标。The present invention tests a large number of face images during debugging, and finds that the images with clear imaging have a relatively high proportion of high-frequency components. After obtaining fftRo, a suitable interval fftRoMin～fftRoMax is set, and the fftRo that falls within this reasonable interval is taken as the follow-up An evaluation metric for the step.

由前述，计算出的指标包括人脸区域的亮部灰度占比brArRoS，人脸T区的灰度方差variGyT，人脸T区的灰度均值AveGyT，人脸区域的灰度动态范围dyRgGy，人脸区域的二维熵imgEpyT，人脸区域的FFT高频占比fftRo一共6个。From the foregoing, the calculated indicators include the gray proportion of the bright part of the face area brArRoS, the gray scale variance variGyT of the face T area, the gray mean value AveGyT of the face T area, the gray dynamic range dyRgGy of the face area, The two-dimensional entropy imgEpyT of the face area, and the FFT high frequency ratio fftRo of the face area are a total of 6.

在获得6个指标后，同时分别设定6个指标的权重，根据指标和对应的权重得到一个调整参数AW：After obtaining 6 indicators, set the weights of the 6 indicators at the same time, and obtain an adjustment parameter AW according to the indicators and the corresponding weights:

AW＝w1*brArRoS+w2*variGyTNo+w3*AveGyTNo+w4*dyRgGyNo+w5*imgEpyT+w6*fftRo。AW=w1*brArRoS+w2*variGyTNo+w3*AveGyTNo+w4*dyRgGyNo+w5*imgEpyT+w6*fftRo.

其中，将variGyT,AveGyT,dyRgGy进行归一化运算得到variGyTNo、AveGyTNo以及dyRgGyNo，w1，w2，w3，w4，w5，w6分别为人脸区域的亮部灰度占比brArRoS，人脸T区的灰度方差variGyT，人脸T区的灰度均值AveGyT，人脸区域的灰度动态范围dyRgGy，人脸区域的二维熵imgEpyT，人脸区域的FFT高频占比fftRo对应的权重系数，按调各指标对补光的的灵敏度来设权重值，由于人脸T区的灰度均值AveGyT灵敏度适中，因此其对应的权重最大，过于灵敏的指标和过于不灵敏的指标对应的权重都不适合占有较大的权重。Among them, variGyT, AveGyT, dyRgGy are normalized to obtain variGyTNo, AveGyTNo and dyRgGyNo, w1, w2, w3, w4, w5, w6 are the gray proportion of the bright part of the face area brArRoS, the gray of the face T area, respectively Degree variance variGyT, gray mean value AveGyT of face T area, gray dynamic range dyRgGy of face area, two-dimensional entropy imgEpyT of face area, FFT high frequency ratio of face area fftRo corresponding weight coefficient, according to the adjustment The sensitivity of each indicator to the supplementary light is used to set the weight value. Since the sensitivity of the gray mean AveGyT in the T-zone of the face is moderate, its corresponding weight is the largest, and the weights corresponding to the indicators that are too sensitive and those that are too insensitive are not suitable for occupation. larger weight.

例如：在一个实施例中，计算了前述的6个指标，每个指标对应的权重可以设置为w1取0.16，w2取0.08，w3取0.46，w4取0.12，w5取0.17，w6取0.13。又例如，在另一个实施例中，计算brArRoS,variGyT,AveGyT,dyRgGy共4个指标，每个指标对应的权重可以设置为w1取0.1，w2取0.05，w3取0.55，w4取0.3。当然，还可以选择其他指标来获得调整指标AW。For example, in one embodiment, the aforementioned 6 indicators are calculated, and the weight corresponding to each indicator can be set to be 0.16 for w1, 0.08 for w2, 0.46 for w3, 0.12 for w4, 0.17 for w5, and 0.13 for w6. For another example, in another embodiment, a total of 4 indicators, brArRoS, variGyT, AveGyT, and dyRgGy, are calculated, and the weight corresponding to each indicator can be set to be 0.1 for w1, 0.05 for w2, 0.55 for w3, and 0.3 for w4. Of course, other indicators can also be selected to obtain the adjustment indicator AW.

然后，本发明即可根据调整参数AW判断人脸图像是否清晰，具体地，判断AW是否处于预设的阈值范围内，如果是，则说明图像清晰，输出当前帧图像；如果AW没有在预设的阈值范围内，则说明图像不清晰，通过通信接口读出当前逆光补偿参数，进行调整，调整当前逆光补偿参数的方法可以有多种，下面举两个示例进行说明：Then, the present invention can judge whether the face image is clear according to the adjustment parameter AW, specifically, judge whether the AW is within the preset threshold range, if so, the image is clear, and the current frame image is output; if the AW is not in the preset threshold Within the threshold range, it means that the image is not clear. Read the current backlight compensation parameters through the communication interface and adjust them. There are many ways to adjust the current backlight compensation parameters. Two examples are given below:

示例一：Example one:

步骤410：根据调整参数AW与预设的阈值范围的上限H或下限L的差值确定调整数值，使用确定出的调整数值一次调整逆光补偿参数。Step 410: Determine the adjustment value according to the difference between the adjustment parameter AW and the upper limit H or lower limit L of the preset threshold range, and use the determined adjustment value to adjust the backlight compensation parameter once.

例如，若AW大于H，则计算AW-H的值，根据计算得到的值减小逆光补偿参数的数值，逆光补偿参数减小的数值与H-AW的值成一定比例。若AW小于L，则计算L-AW的值，根据计算得到的值增加逆光补偿参数的数值，逆光补偿参数增加的数值与L-AW的值成一定比例。本示例可以一次将逆光补偿参数调整到需要的值，缺点是没有反馈机制，调整结果不精确。For example, if AW is greater than H, the value of AW-H is calculated, and the value of the backlight compensation parameter is reduced according to the calculated value. The reduced value of the backlight compensation parameter is proportional to the value of H-AW. If AW is less than L, the value of L-AW is calculated, and the value of the backlight compensation parameter is increased according to the calculated value. The increased value of the backlight compensation parameter is proportional to the value of L-AW. In this example, the backlight compensation parameters can be adjusted to the required value at one time, but the disadvantage is that there is no feedback mechanism, and the adjustment result is imprecise.

示例二：Example two:

步骤420：根据调整参数小于预设的阈值范围的下限或大于预设的阈值范围的上限的情况将当前逆光补偿参数增加或减小固定数值，然后返回最开始的步骤，重复调整逆光补偿参数，直至调整参数在预设的阈值范围内。Step 420: Increase or decrease the current backlight compensation parameter by a fixed value according to the condition that the adjustment parameter is less than the lower limit of the preset threshold range or greater than the upper limit of the preset threshold range, and then returns to the initial step, and repeatedly adjusts the backlight compensation parameter, until the adjustment parameters are within the preset threshold range.

例如，如果调整指参数AW低于下限参数L，则当前逆光补偿参数BL增1，如果调整指参数AW高于上限参数H，则逆光补偿参数BL减1，并将新的逆光补偿参数BL经通讯端口发送到设备，重新执行一遍本方法的步骤，并计算新的当前AW值，再次与L及H进行比较，得到新的BL值，发送到设备，循环往复进行这个调整，直至满足L＜AW＜H，则调整到位，停止本轮调整。For example, if the adjustment index parameter AW is lower than the lower limit parameter L, the current backlight compensation parameter BL is increased by 1; if the adjustment index parameter AW is higher than the upper limit parameter H, the backlight compensation parameter BL is decreased by 1, and the new backlight compensation parameter BL is passed through Send the communication port to the device, perform the steps of this method again, and calculate the new current AW value, compare it with L and H again, get the new BL value, send it to the device, and repeat this adjustment until L< AW<H, the adjustment is in place, and the current round of adjustment is stopped.

实施例2：Example 2:

本发明实施例提供了一种人脸识别逆光补偿装置，如图3所示，该装置包括：An embodiment of the present invention provides a face recognition backlight compensation device, as shown in FIG. 3 , the device includes:

采集模块10，用于在当前逆光补偿参数下采集人脸图像，并在采集到的人脸图像上定位出人脸区域。Theacquisition module 10 is used for acquiring a face image under the current backlight compensation parameter, and locating a face region on the acquired face image.

指标计算模块20，用于计算人脸区域的多个指标并确定多个指标的权重，多个指标包括人脸区域灰度指标和/或人脸区域频率域指标。Theindex calculation module 20 is configured to calculate multiple indices of the face area and determine the weights of the multiple indices, and the multiple indices include the face area grayscale index and/or the face area frequency domain index.

调整参数计算模块30，用于根据人脸区域的多个指标以及多个指标的权重计算调整参数。The adjustmentparameter calculation module 30 is configured to calculate the adjustment parameter according to the multiple indicators of the face area and the weights of the multiple indicators.

调整模块40，用于判断调整参数是否在预设的阈值范围内，若是，不调整当前逆光补偿参数，否则，调整当前逆光补偿参数。Theadjustment module 40 is configured to determine whether the adjustment parameters are within the preset threshold range, if so, do not adjust the current backlight compensation parameters, otherwise, adjust the current backlight compensation parameters.

当采集模块检测到人脸图像中存在人脸时，还可以进一步判断采集到的人脸图像是否稳定，此时，该装置还包括：When the collection module detects that there is a face in the face image, it can further judge whether the collected face image is stable. At this time, the device further includes:

稳定判断模块，用于根据采集到的多帧人脸图像的变化判断人脸图像是否稳定，若是，执行指标计算模块，否则，重新执行采集模块。The stability judgment module is used for judging whether the face image is stable according to the changes of the collected multi-frame face images, if so, execute the index calculation module, otherwise, execute the collection module again.

本发明中，稳定判断模块可以有多种形式，可以选取人脸图像的多种特征指标来反映多帧人脸图像的变化，这里给出一个示例，稳定判断模块包括：In the present invention, the stability judging module can have various forms, and can select various feature indicators of the face image to reflect the changes of the multi-frame face images. Here is an example, the stability judging module includes:

环境区域获取单元，用于将人脸图像减去人脸区域，得到环境区域。The environmental area acquisition unit is used for subtracting the face area from the face image to obtain the environmental area.

灰度方差变化值计算单元，用于分别计算两帧人脸图像的人脸区域和环境区域的灰度方差变化值。The gray-scale variance change value calculation unit is used to calculate the gray-scale variance change values of the face area and the environment area of the two frames of face images respectively.

具体的，灰度方差变化值计算单元包括：Specifically, the gray variance change value calculation unit includes:

灰度方差计算单元，用于计算相邻两帧人脸图像的人脸区域(或环境区域)的灰度方差。The grayscale variance calculation unit is used to calculate the grayscale variance of the face area (or the environment area) of two adjacent frames of face images.

相减单元，用于将相邻两帧人脸图像的人脸区域(或环境区域)的灰度方差相减后求绝对值，得到人脸区域(或环境区域)的灰度方差变化值。The subtraction unit is used to subtract the grayscale variance of the face area (or environment area) of two adjacent frames of face images to obtain the absolute value, and obtain the change value of the grayscale variance of the face area (or environment area).

按照多帧人脸图像的采集时间顺序每隔一定帧数计算一次人脸区域(或环境区域)的灰度方差变化值，得到按时间顺序的人脸区域(或环境区域)的多个灰度方差变化值；According to the acquisition time sequence of multiple frames of face images, the gray variance change value of the face area (or environment area) is calculated every certain number of frames, and the multiple gray levels of the face area (or environment area) in time sequence are obtained. variance change value;

其中，人脸区域的亮部灰度占比通过如下单元模块计算得到：Among them, the gray proportion of the bright part of the face area is calculated by the following unit modules:

人脸区域的亮部灰度占比计算单元，用于计算灰度值在[151,220]之间的像素个数占人脸区域总像素个数的占比，得到人脸区域的亮部灰度占比。The calculation unit of the brightness ratio of the bright part of the face area is used to calculate the proportion of the number of pixels whose gray value is between [151, 220] to the total number of pixels in the face area, and obtain the brightness of the bright part of the face area. proportion.

灰度方差计算单元，用于根据灰度方差的定义公式计算人脸区域的灰度方差、背景区域的灰度方差和人脸T区的灰度方差，图像的灰度方差的定义公式如下：The grayscale variance calculation unit is used to calculate the grayscale variance of the face area, the grayscale variance of the background area, and the grayscale variance of the face T area according to the definition formula of the grayscale variance. The definition formula of the grayscale variance of the image is as follows:

...+...+

......+......+

...+...+

(x1，y1)为图像的左下角坐标，(x2，y2)为图像的右上角坐标，AveGy为图像的灰度均值，gray[x,y]为图像坐标(x,y)像素的灰度值，x＝x1,x1+1,x1+2,……,x2，y＝y1,y1+1,y1+2,……,y2；(x1, y1) is the coordinate of the lower left corner of the image, (x2, y2) is the coordinate of the upper right corner of the image, AveGy is the grayscale mean of the image, and gray[x,y] is the grayscale of the pixel at the image coordinate (x,y) value, x=x1, x1+1, x1+2,...,x2, y=y1, y1+1, y1+2,...,y2;

人脸区域的灰度动态范围通过如下单元模块计算得到：The gray dynamic range of the face area is calculated by the following unit modules:

统计单元，用于统计得到人脸区域的有效最小灰度值和有效最大灰度值，其中：灰度值的像素个数占总像素个数的占比达到设定比值的最小灰度值即为有效最小灰度值，灰度值的像素个数占总像素个数的占比达到设定比值的最大灰度值即为有效最大灰度值；The statistical unit is used to obtain the effective minimum gray value and the effective maximum gray value of the face area by statistics, wherein: the ratio of the number of pixels of the gray value to the total number of pixels reaches the minimum gray value of the set ratio, namely It is the effective minimum gray value, and the maximum gray value whose proportion of the number of gray value pixels to the total number of pixels reaches the set ratio is the effective maximum gray value;

计算单元，用于将有效最大灰度值与有效最小灰度值相减即得到人脸区域的灰度动态范围；a calculation unit, used for subtracting the effective maximum gray value and the effective minimum gray value to obtain the gray dynamic range of the face region;

人脸图像的二维熵通过如下单元模块计算得到：The two-dimensional entropy of the face image is calculated by the following unit modules:

一维熵计算单元，用于进行灰度直方图的统计，即统计人脸区域的灰度级的分布情况，记录下来0～255这个区间内每一灰度级的像素出现的个数，nH[i],i＝0,1,2,……,255，以向量的形式记为nH，按式hG[i]＝nH[i]/nSgy算得向量hG，向量hG长度为255，向量hG的元素hG[i]代表灰度级为i的像素在人脸区域中出现的频率(其中nSgy为人脸区域总像素数)，人脸区域的一维熵imgEpyO的计算公式如下：The one-dimensional entropy calculation unit is used for the statistics of the gray histogram, that is, the distribution of the gray levels in the face area, and the number of pixels in each gray level in the range of 0 to 255 is recorded, nH [i],i=0,1,2,...,255, denoted as nH in the form of a vector, the vector hG is calculated according to the formula hG[i]=nH[i]/nSgy, the length of the vector hG is 255, and the vector hG The element hG[i] represents the frequency of the pixel whose gray level is i in the face area (where nSgy is the total number of pixels in the face area). The calculation formula of the one-dimensional entropy imgEpyO of the face area is as follows:

二维熵计算单元，用于记i为中心像素的灰度级(范围0～255)，j为邻域像素的灰度级(范围0～255)，定义一个特征二元组hHT[i，j]记录人脸区域中中心灰度级为i，邻域像素灰度级为j的像素在人脸区域出现的次数，N为图像尺度，按式hGT[i，j]＝hHT[i，j]/N²算得hGT[i，j]为人脸区域中中心灰度级为i，邻域像素灰度级为j的像素在人脸区域出现的频率；则图像二维熵imgEpyT的计算公式如下，Two-dimensional entropy calculation unit, used to denote i as the gray level of the central pixel (range 0 to 255), j as the gray level of the neighboring pixels (range 0 to 255), and define a feature two-tuple hHT[i, j] Record the number of times that the center gray level is i in the face area, and the pixel whose neighborhood pixel gray level is j appears in the face area, N is the image scale, according to the formula hGT[i, j]=hHT[i, j]/N² can be calculated as hGT[i, j] is the frequency of the pixels whose central gray level is i in the face area and the pixel gray level is j in the neighborhood area in the face area; then the calculation formula of the two-dimensional image entropy imgEpyT as follows,

人脸区域的FFT高频占比通过如下单元模块计算得到：The high frequency proportion of FFT in the face area is calculated by the following unit modules:

傅里叶变换单元，用于对人脸区域进行二维离散傅里叶变换，将人脸区域从空间域转换成频率域；The Fourier transform unit is used to perform two-dimensional discrete Fourier transform on the face area, and convert the face area from the spatial domain to the frequency domain;

高频分量统计单元，用于统计频率域的高频分量占比，即为人脸区域的FFT高频占比。The high-frequency component statistics unit is used to count the proportion of high-frequency components in the frequency domain, that is, the high-frequency proportion of the FFT in the face area.

在获得各个指标后，同时分别设定各个指标的权重，根据指标和对应的权重得到一个调整参数。After each indicator is obtained, the weight of each indicator is set at the same time, and an adjustment parameter is obtained according to the indicator and the corresponding weight.

然后，本发明即可根据调整参数判断人脸图像是否清晰，具体地，判断是否处于预设的阈值范围内，如果是，则说明图像清晰，输出当前帧图像；如果调整参数没有在预设的阈值范围内，则说明图像不清晰，通过通信接口读出当前逆光补偿参数，进行调整，调整当前逆光补偿参数的方式可以有多种，下面举两个示例进行说明：Then, the present invention can judge whether the face image is clear according to the adjustment parameters, specifically, judge whether it is within the preset threshold range, if so, it means that the image is clear, and the current frame image is output; if the adjustment parameters are not within the preset threshold Within the threshold range, the image is not clear. Read the current backlight compensation parameters through the communication interface and adjust them. There are many ways to adjust the current backlight compensation parameters. Two examples are given below:

通过如下单元模块调整当前逆光补偿参数：Adjust the current backlight compensation parameters through the following unit modules:

示例一：Example one:

第一调整单元，用于根据调整参数与预设的阈值范围的上限或下限的差值确定调整数值，使用确定出的调整数值一次调整逆光补偿参数。The first adjustment unit is configured to determine the adjustment value according to the difference between the adjustment parameter and the upper limit or lower limit of the preset threshold range, and use the determined adjustment value to adjust the backlight compensation parameter at one time.

示例二：Example two:

第二调整单元，用于根据调整参数小于预设的阈值范围的下限或大于预设的阈值范围的上限的情况将当前逆光补偿参数增加或减小固定数值，然后返回最开始的采集模块，重复调整逆光补偿参数，直至调整参数在预设的阈值范围内。The second adjustment unit is configured to increase or decrease the current backlight compensation parameter by a fixed value according to the condition that the adjustment parameter is smaller than the lower limit of the preset threshold range or larger than the upper limit of the preset threshold range, and then returns to the initial acquisition module, and repeats Adjust the backlight compensation parameters until the adjustment parameters are within the preset threshold range.

本发明实施例所提供的装置，其实现原理及产生的技术效果和前述方法实施例相同，为简要描述，装置实施例部分未提及之处，可参考前述方法实施例中相应内容。所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，前述描述的装置和单元的具体工作过程，均可以参考上述方法实施例中的对应过程，在此不再赘述。The implementation principle and technical effects of the device provided by the embodiment of the present invention are the same as those of the foregoing method embodiment. For brief description, for the parts not mentioned in the device embodiment, reference may be made to the corresponding content in the foregoing method embodiment. Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the apparatuses and units described above can refer to the corresponding processes in the above method embodiments, and details are not repeated here.

实施例3：Example 3:

本说明书提供的上述实施例所述的方法或装置可以通过计算机程序实现业务逻辑并记录在存储介质上，所述的存储介质可以计算机读取并执行，实现本说明书实施例所描述方案的效果。因此，本发明还提供用于人脸识别逆光补偿的计算机可读存储介质，包括处理器及用于存储处理器可执行指令的存储器，指令被处理器执行时实现包括实施例1的人脸识别逆光补偿方法的步骤。The methods or apparatuses described in the above embodiments provided in this specification can realize business logic through computer programs and record them on a storage medium, and the storage medium can be read and executed by a computer to achieve the effects of the solutions described in the embodiments of this specification. Therefore, the present invention also provides a computer-readable storage medium for face recognition backlight compensation, including a processor and a memory for storing processor-executable instructions, and when the instructions are executed by the processor, the face recognition including Embodiment 1 is implemented. The steps of the backlight compensation method.

所述存储介质可以包括用于存储信息的物理装置，通常是将信息数字化后再以利用电、磁或者光学等方式的媒体加以存储。所述存储介质有可以包括：利用电能方式存储信息的装置如，各式存储器，如RAM、ROM等；利用磁能方式存储信息的装置如，硬盘、软盘、磁带、磁芯存储器、磁泡存储器、U盘；利用光学方式存储信息的装置如，CD或DVD。当然，还有其他方式的可读存储介质，例如量子存储器、石墨烯存储器等等。The storage medium may include a physical device for storing information, and usually the information is digitized and then stored in an electrical, magnetic or optical medium. The storage medium may include: devices that use electrical energy to store information, such as various memories, such as RAM, ROM, etc.; devices that use magnetic energy to store information, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, magnetic bubble memories, etc. USB stick; a device that stores information optically, such as a CD or DVD. Of course, there are other readable storage media, such as quantum memory, graphene memory, and so on.

上述所述的装置根据方法实施例的描述还可以包括其他的实施方式。具体的实现方式可以参照相关方法实施例的描述，在此不作一一赘述。The above-mentioned apparatus may further include other implementation manners according to the description of the method embodiment. For a specific implementation manner, reference may be made to the description of the related method embodiments, which will not be repeated here.

实施例4：Example 4:

本发明还提供一种用于人脸识别逆光补偿的设备，所述的设备可以为单独的计算机，也可以包括使用了本说明书的一个或多个所述方法或一个或多个实施例装置的实际操作装置等。所述用于人脸识别逆光补偿的设备可以包括至少一个处理器以及存储计算机可执行指令的存储器，处理器执行所述指令时实现上述任意一个或者多个实施例中所述人脸识别逆光补偿方法的步骤。The present invention also provides a device for face recognition backlight compensation, the device may be a separate computer, or may include a device using one or more of the methods or one or more embodiments of this specification. Actual operating device, etc. The device for face recognition backlight compensation may include at least one processor and a memory for storing computer-executable instructions. When the processor executes the instructions, the face recognition backlight compensation described in any one or more of the above embodiments is implemented. steps of the method.

上述所述的设备根据方法或者装置实施例的描述还可以包括其他的实施方式，具体的实现方式可以参照相关方法实施例的描述，在此不作一一赘述。The above-mentioned device may also include other implementation manners according to the description of the method or apparatus embodiment, and the specific implementation manner may refer to the description of the related method embodiment, which will not be repeated here.

需要说明的是，本说明书上述所述的装置或者系统根据相关方法实施例的描述还可以包括其他的实施方式，具体的实现方式可以参照方法实施例的描述，在此不作一一赘述。本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于硬件+程序类、存储介质+程序实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。It should be noted that the above-mentioned apparatus or system in this specification may also include other implementation manners according to the description of the related method embodiments, and the specific implementation manner may refer to the description of the method embodiments, which will not be repeated here. Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the various embodiments may be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the hardware+program class, storage medium+program embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant part may refer to the partial description of the method embodiment.

上述对本说明书特定实施例进行了描述。其它实施例在所附权利要求书的范围内。在一些情况下，在权利要求书中记载的动作或步骤可以按照不同于实施例中的顺序来执行并且仍然可以实现期望的结果。另外，在附图中描绘的过程不一定要求示出的特定顺序或者连续顺序才能实现期望的结果。在某些实施方式中，多任务处理和并行处理也是可以的或者可能是有利的。The foregoing describes specific embodiments of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that in the embodiments and still achieve desirable results. Additionally, the processes depicted in the figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

上述实施例阐明的系统、装置、模块或单元，具体可以由计算机芯片或实体实现，或者由具有某种功能的产品来实现。一种典型的实现设备为计算机。具体的，计算机例如可以为个人计算机、膝上型计算机、车载人机交互设备、蜂窝电话、相机电话、智能电话、个人数字助理、媒体播放器、导航设备、电子邮件设备、游戏控制台、平板计算机、可穿戴设备或者这些设备中的任何设备的组合。The systems, devices, modules or units described in the above embodiments may be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, an in-vehicle human-computer interaction device, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet A computer, wearable device, or a combination of any of these devices.

为了描述的方便，描述以上装置时以功能分为各种模块分别描述。当然，在实施本说明书一个或多个时可以把各模块的功能在同一个或多个软件和/或硬件中实现，也可以将实现同一功能的模块由多个子模块或子单元的组合实现等。以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。For the convenience of description, when describing the above device, the functions are divided into various modules and described respectively. Of course, when implementing one or more of this specification, the functions of each module can be implemented in the same one or more software and/or hardware, and the modules that implement the same function can also be implemented by a combination of multiple sub-modules or sub-units, etc. . The apparatus embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated. to another system, or some features can be ignored, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.

本领域技术人员也知道，除了以纯计算机可读程序代码方式实现控制器以外，完全可以通过将方法步骤进行逻辑编程来使得控制器以逻辑门、开关、专用集成电路、可编程逻辑控制器和嵌入微控制器等的形式来实现相同功能。因此这种控制器可以被认为是一种硬件部件，而对其内部包括的用于实现各种功能的装置也可以视为硬件部件内的结构。或者甚至，可以将用于实现各种功能的装置视为既可以是实现方法的软件模块又可以是硬件部件内的结构。Those skilled in the art also know that, in addition to implementing the controller in the form of pure computer-readable program code, the controller can be implemented as logic gates, switches, application-specific integrated circuits, programmable logic controllers and embedded devices by logically programming the method steps. The same function can be realized in the form of a microcontroller, etc. Therefore, such a controller can be regarded as a hardware component, and the devices included therein for realizing various functions can also be regarded as a structure within the hardware component. Or even, the means for implementing various functions can be regarded as both a software module implementing a method and a structure within a hardware component.

本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block in the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in a flow or flow of a flowchart and/or a block or blocks of a block diagram.

这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions The apparatus implements the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。These computer program instructions can also be loaded on a computer or other programmable data processing device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the flow or blocks of the flowcharts and/or the block or blocks of the block diagrams.

在一个典型的配置中，计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

还需要说明的是，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、商品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、商品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法或者设备中还存在另外的相同要素。It should also be noted that the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device comprising a series of elements includes not only those elements, but also Other elements not expressly listed, or which are inherent to such a process, method, article of manufacture, or apparatus are also included. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in the process, method, or device that includes the element.

本领域技术人员应明白，本说明书一个或多个实施例可提供为方法、系统或计算机程序产品。因此，本说明书一个或多个实施例可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本说明书一个或多个实施例可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。As will be appreciated by one skilled in the art, one or more embodiments of this specification may be provided as a method, system or computer program product. Accordingly, one or more embodiments of this specification may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of the present specification may employ a computer program implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein form of the product.

本说明书一个或多个实施例可以在由计算机执行的计算机可执行指令的一般上下文中描述，例如程序模块。一般地，程序模块包括执行特定任务或实现特定抽象数据类型的例程、程序、对象、组件、数据结构等等。也可以在分布式计算环境中实践本本说明书一个或多个实施例，在这些分布式计算环境中，由通过通信网络而被连接的远程处理设备来执行任务。在分布式计算环境中，程序模块可以位于包括存储设备在内的本地和远程计算机存储介质中。One or more embodiments of this specification may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. One or more embodiments of the present specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including storage devices.

本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于系统实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。在本说明书的描述中，参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含于本说明书的至少一个实施例或示例中。在本说明书中，对上述术语的示意性表述并不必须针对的是相同的实施例或示例。而且，描述的具体特征、结构、材料或者特点可以在任一个或多个实施例或示例中以合适的方式结合。此外，在不相互矛盾的情况下，本领域的技术人员可以将本说明书中描述的不同实施例或示例以及不同实施例或示例的特征进行结合和组合。Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the various embodiments may be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, as for the system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for related parts, please refer to the partial descriptions of the method embodiments. In the description of this specification, description with reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples", etc., mean specific features described in connection with the embodiment or example , structure, material or feature is included in at least one embodiment or example of this specification. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, those skilled in the art may combine and combine the different embodiments or examples described in this specification, as well as the features of the different embodiments or examples, without conflicting each other.

最后应说明的是：以上所述实施例，仅为本发明的具体实施方式，用以说明本发明的技术方案，而非对其限制，本发明的保护范围并不局限于此，尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：任何熟悉本技术领域的技术人员在本发明揭露的技术范围内，其依然可以对前述实施例所记载的技术方案进行修改或可轻易想到变化，或者对其中部分技术特征进行等同替换；而这些修改、变化或者替换，并不使相应技术方案的本质脱离本发明实施例技术方案的精神和范围。都应涵盖在本发明的保护范围之内。因此，本发明的保护范围应以所述权利要求的保护范围为准。Finally, it should be noted that the above-mentioned embodiments are only specific implementations of the present invention, and are used to illustrate the technical solutions of the present invention, but not to limit them. The protection scope of the present invention is not limited thereto, although referring to the foregoing The embodiment has been described in detail the present invention, those of ordinary skill in the art should understand: any person skilled in the art who is familiar with the technical field within the technical scope disclosed by the present invention can still modify the technical solutions described in the foregoing embodiments. Changes can be easily conceived, or equivalent replacements are made to some of the technical features; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. All should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

Translated fromChinese

1.一种人脸识别逆光补偿方法，其特征在于，所述方法包括：1. a face recognition backlight compensation method, is characterized in that, described method comprises:

2.根据权利要求1所述的人脸识别逆光补偿方法，其特征在于，在当前逆光补偿参数下采集到的人脸图像为多帧，所述方法还包括：2. The face recognition backlight compensation method according to claim 1, wherein the face images collected under the current backlight compensation parameters are multiple frames, and the method further comprises:

3.根据权利要求2所述的人脸识别逆光补偿方法，其特征在于，所述根据采集到的多帧人脸图像的变化判断人脸图像是否稳定包括：3. The face recognition backlight compensation method according to claim 2, wherein the judging whether the face image is stable according to the change of the collected multi-frame face images comprises:

4.根据权利要求3所述的人脸识别逆光补偿方法，其特征在于，所述分别计算两帧人脸图像的人脸区域和环境区域的灰度方差变化值包括：4. face recognition backlight compensation method according to claim 3, is characterized in that, described calculating respectively the face area of two frames of face images and the gray level variance change value of environment area comprise:

5.根据权利要求4所述的人脸识别逆光补偿方法，其特征在于，所述人脸区域灰度指标包括人脸区域的亮部灰度占比、人脸T区的灰度方差、人脸T区的灰度均值、人脸区域的灰度动态范围、人脸区域的二维熵，所述人脸区域频率域指标包括人脸区域的FFT高频占比，所述多个指标包括人脸区域的亮部灰度占比、人脸T区的灰度方差、人脸区域的灰度动态范围、人脸区域的二维熵、人脸区域的FFT高频占比中的至少一个以及人脸T区的灰度均值。5. The face recognition backlight compensation method according to claim 4, wherein the gray scale index of the face region comprises the gray scale ratio of the bright part of the face region, the gray scale variance of the T region of the human face, the The gray mean value of the face T area, the gray dynamic range of the face area, and the two-dimensional entropy of the face area, the face area frequency domain index includes the FFT high frequency ratio of the face area, and the multiple indexes include At least one of the grayscale proportion of the bright part of the face region, the grayscale variance of the face T zone, the grayscale dynamic range of the face region, the two-dimensional entropy of the face region, and the high frequency ratio of the FFT of the face region And the gray mean of the T-zone of the face.

6.根据权利要求1-5任一所述的人脸识别逆光补偿方法，其特征在于，调整当前逆光补偿参数的方法为：6. The face recognition backlight compensation method according to any one of claims 1-5, wherein the method for adjusting the current backlight compensation parameter is:

7.一种人脸识别逆光补偿装置，其特征在于，所述装置包括：7. A face recognition backlight compensation device, wherein the device comprises:

指标计算模块，用于计算人脸区域的多个指标并确定多个指标的权重，所述多个指标包括人脸区域灰度指标和/或人脸区域频率域指标；an index calculation module for calculating multiple indices of the face area and determining the weights of the multiple indices, the multiple indices including the face area grayscale index and/or the face area frequency domain index;

8.根据权利要求7所述的人脸识别逆光补偿装置，其特征在于，在当前逆光补偿参数下采集到的人脸图像为多帧，所述装置还包括：8. The face recognition backlight compensation device according to claim 7, wherein the collected face images under the current backlight compensation parameters are multiple frames, and the device further comprises:

稳定判断模块，用于根据采集到的多帧人脸图像的变化判断人脸图像是否稳定，若是，执行所述指标计算模块，否则，重新执行所述采集模块；a stability judging module, configured to judge whether the face image is stable according to the changes of the collected multi-frame face images, if so, execute the index calculation module, otherwise, execute the collection module again;

所述稳定判断模块包括：The stability judgment module includes:

9.一种用于人脸识别逆光补偿的计算机可读存储介质，其特征在于，包括处理器及用于存储处理器可执行指令的存储器，所述指令被所述处理器执行时实现包括权利要求1-6任一所述人脸识别逆光补偿方法的步骤。9. A computer-readable storage medium for face recognition backlight compensation, characterized in that it comprises a processor and a memory for storing processor-executable instructions, and when the instructions are executed by the processor, the Steps of any one of the face recognition backlight compensation methods are required in 1-6.

10.一种用于人脸识别逆光补偿的设备，其特征在于，包括至少一个处理器以及存储计算机可执行指令的存储器，所述处理器执行所述指令时实现权利要求1-6中任意一项所述人脸识别逆光补偿方法的步骤。10. A device for face recognition backlight compensation, characterized in that it comprises at least one processor and a memory for storing computer-executable instructions, which implements any one of claims 1-6 when the processor executes the instructions The steps of the face recognition backlight compensation method described in item 1.